基于局部敏感哈希的近邻传播聚类

本文针对近邻传播聚类中存在的复杂度高问题,提出了局部敏感哈希的近邻传播聚类算法,根据局部敏感哈希先将相似数据哈希到同一桶中,在对每个桶中的数据进行聚类。实验结果表明,该算法降低了复杂度,提高了准确率。     

基于局部敏感哈希的隐私保护实时服务推荐

协同过滤算法是服务推荐系统中最有效和应用最广泛的推荐方法,其侧重于提高推荐结果的准确性。然而,在大数据背景下,用户行为数据不仅经常频繁更新而且数据规模增长迅速,传统的协同过滤算法需要穷举搜索所有数据,相似度计算耗时较高,推荐效率低,无法满足用户实时体验的需求服务。快速从大数据中获得高质量的推荐服务成为一种新的需求,为此,提出基于局部敏感哈希技术的协同过滤算法,算法过滤了绝大多数不相似的项目,避免了冗余的相似度计算,另一方面算法将用户行为数据哈希为二进制哈希编码,进而保护用户隐私。最后,在不同规模尺寸的数据集上与主流算法对比,实验表明提出的算法在效率和准确度间能够取得较好的折衷。     

开放环境多分布特性的局部敏感哈希检索方法

基于局部敏感哈希的检索方法能够较好地解决高维大规模数据的近似近邻检索问题.但在开放环境下针对多种分布特性时,迄今尚未有令人满意的解决方案.利用Laplacian算子对数据分布剧烈变化敏感的特性,提出一种具有全局性、适用于开放环境下多种分布特性的基于Laplacian算子的局部敏感哈希搜索方法(LPLSH).该方法把La...     

基于位置信息熵的局部敏感哈希聚类方法

在分析海量生物序列时,现有的聚类算法存在着时间效率不高、准确率较低,以及聚类结果的生物意义不足等问题。针对这些问题,提出一种基于位置信息熵的局部敏感哈希聚类方法。通过对生物序列使用K词计算其标准熵,将标准熵作为局部敏感哈希函数簇的特征向量,计算特征矩阵并应用于生物序列聚类。实验结果表明,该算法能够有效地提高时间效率和聚类的准确率。随着数据集的增大,也同样取得很好的效果,实验结果更具有生物解释性和实际意义。     

基于余弦距离的局部敏感哈希的KNN算法在中文文本上的快速分类

文本分类是文本挖掘中最重要的研究内容之一。为了克服目前以距离衡量的近似分类算法在海量数据下耗费大量时间的缺陷,提出了结合基于余弦距离的局部敏感哈希的方式将KNN算法在TF-IDF下对中文文本进行快速分类。同时结合文本数据的特性给出了不同的哈希函数级联方式分别进行实验。在实验过程采用了布尔向量的方式规避重复访问,使分类的结果在可以允许的范围内,分类速度比原始KNN提高了许多。     

基于改进的局部敏感哈希算法实现图像型垃圾邮件过滤

提出一种快速的图像型垃圾邮件过滤方案,结合半监督机器学习技术改进局部敏感哈希(LSH)算法,基于改进的LSH算法构建垃圾图像特征库索引,提高图像的查找速度。构造了60000个垃圾图像样本,实验结果表明利用改进的LSH算法能有效地提高垃圾图像的过滤速度。     

面向视觉搜索的空间局部敏感哈希方法

目的 视觉检索需要准确、高效地从大型图像或者视频数据集中检索出最相关的视觉内容,但是由于数据集中图像数据量大、特征维度高的特点,现有方法很难同时保证快速的检索速度和较好的检索效果。方法 对于面向图像视频数据的高维数据视觉检索任务,提出加权语义局部敏感哈希算法（weighted semantic locality-sensitive hashing, WSLSH）。该算法利用两层视觉词典对参考特征空间进行二次空间划分,在每个子空间里使用加权语义局部敏感哈希对特征进行精确索引。其次,设计动态变长哈希码,在保证检索性能的基础上减少哈希表数量。此外,针对局部敏感哈希（locality sensitive hashing, LSH）的随机不稳定性,在LSH函数中加入反映参考特征空间语义的统计性数据,设计了一个简单投影语义哈希函数以确保算法检索性能的稳定性。结果 在Holidays、Oxford5k和DataSetB数据集上的实验表明,WSLSH在DataSetB上取得最短平均检索时间0.034 25 s;在编码长度为64位的情况下,WSLSH算法在3个数据集上的平均精确度均值（mean average precision,mAP）分别提高了1.2%32.6%、1.7%19.1%和2.6%28.6%,与几种较新的无监督哈希方法相比有一定的优势。结论 通过进行二次空间划分、对参考特征的哈希索引次数进行加权、动态使用变长哈希码以及提出简单投影语义哈希函数来对LSH算法进行改进。由此提出的加权语义局部敏感哈希（WSLSH）算法相比现有工作有更快的检索速度,同时,在长编码的情况下,取得了更为优异的性能。     

区块链环境中基于局部敏感哈希的协同过滤推荐研究

针对区块链环境中海量高维的数据使得推荐性能低下的问题,通过对局部敏感哈希算法的优化,降低其在近邻搜索过程中带来的额外计算和存储开销.利用数据分布的主成分减少传统LSH中不良捕获的投影方向,同时对投影向量权重进行量化,以减少哈希表和哈希函数的使用;通过对哈希桶的间隔进行调整,并且根据冲突次数的大小进一步细化查询结果集,以...     

Efficient combination of index tables and hashing

It is shown that for data stored on direct access devices access time can be reduced without increasing storage demand, through a single level index table which itself is accessible by hashing. If the complete index table can be stored in main memory this method is always superior to direct hashing and to sequentially organized index tables. If the index table is stored on disk it always yields smaller access time than multi-level index tables and—depending on the size of the index table and on the number of records per track—it is comparable or better than hashing the data directly. Expressions are given to determine in this case which method is more efficient.     

Leader-following control of multiple nonholonomic systems over directed communication graphs

This paper considers the leader-following control problem of multiple nonlinear systems with directed communication topology and a leader. If the state of each system is measurable, distributed state feedback controllers are proposed using neighbours’ state information with the aid of Lyapunov techniques and properties of Laplacian matrix for time-invariant communication graph and time-varying communication graph. It is shown that the state of each system exponentially converges to the state of a leader. If the state of each system is not measurable, distributed observer-based output feedback control laws are proposed. As an application of the proposed results, formation control of wheeled mobile robots is studied. The simulation results show the effectiveness of the proposed results.     

Code completion of multiple keywords from abbreviated input

Abbreviation Completion is a novel technique to improve the efficiency of code-writing by supporting code completion of multiple keywords based on non-predefined abbreviated input??a different approach from conventional code completion that finds one keyword at a time based on an exact character match. Abbreviated input consisting of abbreviated keywords and non-alphanumeric characters between each abbreviated keyword (e.g. pb st nm) is expanded into a full expression (e.g. public String name) by a Hidden Markov Model learned from a corpus of existing code and abbreviation examples. The technique does not require the user to memorize abbreviations and provides incremental feedback of the most likely completions. In addition to code completion by disabbreviation of multiple keywords, abbreviation completion supports prediction of the next keywords and non-alphanumeric characters of a code completion candidate, a technique called code completion by extrapolation. The system finds the most likely next keywords and non-alphanumeric characters using an n-gram model of programming language. This enables a code completion scenario in which a user first types a short abbreviated expression to complete the beginning part of a desired full expression and then uses the extrapolation feature to complete the remaining part without further typing. This paper presents the algorithm for abbreviation completion, integrated with a new user interface for multiple-keyword code completion. We tested the system by sampling 4919 code lines from open source projects and found that more than 99% of the code lines could be resolved from acronym-like abbreviations. The system could also extrapolate code completion candidates to complete the next one or two keywords with the accuracy of 96% and 82%, respectively. A user study of code completion by disabbreviation found 30% reduction in time usage and 41% reduction of keystrokes over conventional code completion.     

A robust alignment-free fingerprint hashing algorithm based on minimum distance graphs

Abstraction of a fingerprint in the form of a hash can be used for secure authentication. The main challenge is in finding the right choice of features which remain relatively invariant to distortions such as rotation, translation and minutiae insertions and deletions, while at the same time capturing the diversity across users. In this paper, an alignment-free novel fingerprint hashing algorithm is proposed which uses a graph comprising of the inter-minutia minimum distance vectors originating from the core point as a feature set called the minimum distance graph. Matching of hashes has been implemented using a corresponding search algorithm. Based on the experiments conducted on the FVC2002-DB1a and FVC2002-DB2a databases, we obtained an equal error rate of 2.27%. The computational cost associated with our fingerprint hash generation and matching processes is relatively low, despite its success in capturing the minutia positional variations across users.     

Hidden semantic hashing for fast retrieval over large scale document collection

Multimedia Tools and Applications - As is well known, the semantics of documents are exposed to us in latent way. However, most existing hashing methods ignore this fact and thus fail to discover...     

面向视频语义分析的局部敏感的可鉴别稀疏表示

视频语义分析已经成为人们研究的热点。在传统稀疏表示方法中,相似视频特征未必能产生相近稀疏表示结果。在基于稀疏表示的视频语义分析中,假定相似的视频数据样本的稀疏表示也相似,即两个相似视频特征的稀疏系数之间的距离较小。为了提高视频语义分析的准确性,基于该假设提出一种面向视频语义分析的局部敏感的可鉴别稀疏表示方法。该方法在局部敏感稀疏表示中引入基于稀疏系数的鉴别损失函数,优化构建稀疏表示的字典,使稀疏表示特征满足类内离散度小、类间离散度大的Fisher准则,并建立可鉴别稀疏模型。为验证所提方法的有效性,在相关视频数据库中将其与多种算法进行对比,实验结果表明,该方法显著地提高了视频特征稀疏表示的鉴别性,有效地提高了视频语义分析的准确性。     

基于相对距离哈希方法的一种高维索引

人们设计了许多索引以有效地处理高维空间中的近邻查询和区域查询。已经证明,维数较高时利用高维索引处理这两类查询几乎不可能比线性扫描快。提出了一种两层索引以自适应地识别数据集中的聚簇;数据集具有聚簇特性时,用该索引处理邻近查询和区域查询比现有的索引结构快;对其他数据集,利用该索引处理邻近查询和区域查询与线性扫描大致相当。该索引的上层结构将一些参考点组织成一棵二叉树,下层结构是一系列动态哈希表。数据集中的数据点根据它们到参考点的相对距离被哈希到相应的哈希桶中。查询处理时用查询点到参考点的距离进行剪除搜索。实验表明,提出的索引结构具有良好的性能。